Getting Started with Matplotlib

Level: Beginner

Table of Contents

1. Course Overview
2. Understanding Matplotlib Architecture
   • Installation
   • Figure Anatomy
   • Three-Layer Architecture
   • Backends
   • Demo: Creating Figures
   • Text and Colors
   • Demo: Subplots
3. Building Line Charts
   • When to Use a Line Chart
   • Dataset: Yellow Taxi Trips NYC
   • Demo: Creating Line Charts
   • Demo: Customizing Line Charts
4. Working with Scatter Plots
   • When to Use a Scatter Plot
   • Demo: Creating Scatter Plots
   • Demo: Customizing Scatter Plots
5. Creating Bar Charts
   • When to Use a Bar Chart
   • Demo: Creating Bar Charts
   • Demo: Customizing Bar Charts
   • Demo: Exporting Figures
6. Reference Tables
7. Reference Diagrams

1. Course Overview

Matplotlib is a data visualization library for Python. This course covers the basics for creating and customizing fundamental charts: line charts, bar charts, and scatter plots.

Prerequisites:

• Familiarity with the Jupyter Notebook interface
• Data manipulation with Pandas

What you will learn:

• Matplotlib architecture (Figure, Axes, Artist)
• Both APIs: scripting layer (plt.*) and artist layer (OOP)
• Creating and customizing line charts, scatter plots, and bar charts
• Exporting figures to PNG/PDF

2. Understanding Matplotlib Architecture

2.1 Installation

# Via pip
pip install matplotlib

# Via conda (Anaconda distribution)
conda install matplotlib

Verify installed version:

import matplotlib
print(matplotlib.__version__)  # e.g.: 3.6.2

2.2 Figure Anatomy

Matplotlib has its own vocabulary. Understanding these terms from the start makes reading documentation and debugging code easier.

Object hierarchy:

Figure
  └── Axes (1 or more)
        ├── XAxis
        │     ├── tick marks (major/minor)
        │     └── tick labels
        ├── YAxis
        │     ├── tick marks (major/minor)
        │     └── tick labels
        ├── Title
        ├── Line2D
        ├── Text (annotations, labels)
        └── Patch (rectangles for bar charts, etc.)

2.3 Three-Layer Architecture

Matplotlib’s architecture is conceptualized as a stack of three layers:

┌─────────────────────────────────────────┐
│  Scripting Layer  (pyplot — plt.*)      │  ← Simple user interface
│  High-level API, implicit               │
├─────────────────────────────────────────┤
│  Artist Layer  (OOP / explicit API)     │  ← Fine control over each element
│  Figure, Axes, Line2D, Text, Patch…     │
├─────────────────────────────────────────┤
│  Backend Layer                          │  ← Actual rendering in environment
│  (inline, notebook, Qt5Agg, Agg…)       │
└─────────────────────────────────────────┘

2.4 Backends

The backend determines how and where the figure is displayed. Matplotlib automatically chooses the appropriate backend.

# Check active backend
import matplotlib.pyplot as plt
plt.get_backend()  # 'module://matplotlib_inline.backend_inline' in Jupyter

# Enable interactive backend in Jupyter Notebook
import ipympl
%matplotlib notebook

Note: In Jupyter Notebook, the inline backend automatically calls plt.show() at the end of each cell. In other environments, you must call plt.show() explicitly.

2.5 Demo: Creating Figures

Scripting Layer (Implicit API)

import matplotlib.pyplot as plt
import numpy as np

# Plot a single point
plt.plot(2, 2, '.')

# Plot a simple curve
x = np.array([1, 2, 3, 4, 5])
y = x * 10
plt.plot(x, y)
plt.show()

Artist Layer (Explicit API / OOP)

import matplotlib.pyplot as plt
import numpy as np

x = np.array([1, 2, 3, 4, 5])
y = x * 10

# Create Figure and Axes separately
fig = plt.figure()
ax = plt.subplot()
ax.plot(x, y)
plt.show()

Difference Between the Two APIs

2.6 Text and Colors

Adding Text

Text Customization Parameters

plt.title(
    'Basic plot',
    family='serif',          # font family: serif, monospace, sans, cursive, fantasy
    color='#00A3FF',         # color (hex, name, letter)
    weight='bold',           # weight: bold, normal, black, ultralight (or 0–1000)
    size=20,                 # size in points or: small, medium, large
    ha='left',               # horizontal alignment: left, center, right
    x=0                      # relative x position (0=left, 1=right)
)

Color Specification

# Short letter
plt.plot(x, y, color='r')          # red

# Full name
plt.plot(x, y, color='green')

# Hexadecimal code
plt.plot(x, y, color='#00A3FF')

# RGB tuple (0–1)
plt.plot(x, y, color=(0.1, 0.5, 0.9))

2.7 Demo: Subplots

A Figure can contain multiple Axes. They are positioned using a grid system (rows × columns).

import matplotlib.pyplot as plt
import numpy as np

a = np.random.randint(0, high=50, size=5)
b = np.random.randint(0, high=50, size=5)
x = np.random.randint(0, high=50, size=5)
y = np.random.randint(0, high=50, size=5)

# 2×2 grid: 4 subplots
fig = plt.figure()
ax1 = plt.subplot(2, 2, 1)   # row 1, column 1
ax2 = plt.subplot(2, 2, 2)   # row 1, column 2
ax3 = plt.subplot(2, 2, 3)   # row 2, column 1
ax4 = plt.subplot(2, 2, 4)   # row 2, column 2

ax1.plot(a, b, '.')
ax2.plot(a, x, '.')
ax3.plot(x, y, '.')
ax4.plot(b, y, '.')

plt.show()

Alternative syntax with plt.subplots() (recommended):

fig, axes = plt.subplots(nrows=2, ncols=2, figsize=(10, 8))

axes[0, 0].plot(a, b, '.')
axes[0, 1].plot(a, x, '.')
axes[1, 0].plot(x, y, '.')
axes[1, 1].plot(b, y, '.')

plt.tight_layout()  # automatic spacing adjustment
plt.show()

3. Building Line Charts

3.1 When to Use a Line Chart

Line charts are ideal for:

• Showing trends over time
• Identifying outliers in a time series
• Making predictions about future evolution
• Comparing multiple series on the same time scale

3.2 Dataset: Yellow Taxi Trips NYC

The dataset used throughout this course is a public dataset of Yellow Taxi trips in New York City (January 2022, source: NYC Government).

Main columns:

3.3 Demo: Creating Line Charts

Data Preparation

import matplotlib.pyplot as plt
import pandas as pd

# Import 5 CSV files (one per week)
df1 = pd.read_csv('Yellow Taxi Trips - Week 1.csv')
df2 = pd.read_csv('Yellow Taxi Trips - Week 2.csv')
df3 = pd.read_csv('Yellow Taxi Trips - Week 3.csv')
df4 = pd.read_csv('Yellow Taxi Trips - Week 4.csv')
df5 = pd.read_csv('Yellow Taxi Trips - Week 5.csv')

full_dataset = pd.concat([df1, df2, df3, df4, df5])

# Convert date column to proper type
full_dataset['Pickup_Date'] = pd.to_datetime(
    full_dataset['Pickup_Date'],
    format='%d/%m/%Y'
)

# Aggregation: total trips per day
trips = full_dataset.groupby(
    by='Pickup_Date',
    as_index=False
)['ID'].count()

trips.rename(columns={'ID': 'Trips'}, inplace=True)

Creating the Line Chart

# --- Scripting Layer (implicit) ---
plt.plot(trips['Pickup_Date'], trips['Trips'])
plt.show()

# --- Artist Layer (explicit / OOP) ---
fig, ax = plt.subplots()
ax.plot(trips['Pickup_Date'], trips['Trips'])
plt.show()

3.4 Demo: Customizing Line Charts

fig, ax = plt.subplots(figsize=(9, 4))

ax.plot(
    trips['Pickup_Date'],
    trips['Trips'],
    '#5A5A5A',           # line color
    linewidth=2,          # line thickness
    linestyle='--',       # style: '-', '--', '-.', ':'
    marker='o',           # marker type
    markersize=5,         # marker size
    markerfacecolor='blue'
)

ax.set_title('Yellow Taxi Trips in New York City')
ax.set_xlabel('Pickup Date')
ax.set_ylabel('Trips')
ax.set_ylim(0, 120000)

fig.autofmt_xdate()  # auto-rotate date labels

plt.show()

plt.plot() / ax.plot() Key Parameters

4. Working with Scatter Plots

4.1 When to Use a Scatter Plot

Scatter plots are used to:

• Visualize the relationship between two variables
• Identify the direction (positive/negative) and strength of a correlation
• Detect clusters and outliers
• Correlation analysis (without implying causality)

⚠️ Warning: Correlation ≠ Causation.

Overplotting solution: When many points overlap, use:

• marker='.' — smaller marker type
• s=5 — reduce marker size
• alpha=0.15 — adjust transparency

4.2 Demo: Creating Scatter Plots

import matplotlib.pyplot as plt
import pandas as pd

taxi_trips = pd.read_csv('Yellow Taxi Trips - Week 1.csv')

JFK_trips = taxi_trips[
    (taxi_trips['Payment_Type'] == 'Credit card') &
    (taxi_trips['Rate_Code'] == 'JFK') &
    (taxi_trips['Passenger_Count'] > 0) &
    (taxi_trips['Total_Amount'] < 90)
]

# Method 1: plt.scatter()
plt.scatter(JFK_trips['Total_Amount'], JFK_trips['Tip_Amount'])
plt.show()

# Method 2: plt.plot() with linestyle='None'
plt.plot(
    JFK_trips['Total_Amount'],
    JFK_trips['Tip_Amount'],
    linestyle='None',
    marker='o'
)
plt.show()

4.3 Demo: Customizing Scatter Plots

Scatter plot with colorbar (3rd variable encoded)

plt.figure(figsize=(9, 4))

scatter = plt.scatter(
    JFK_trips['Total_Amount'],
    JFK_trips['Tip_Amount'],
    marker='o',
    s=10,
    c=JFK_trips['Passenger_Count'],    # color encoded by 3rd variable
    cmap='plasma',                     # colormap
    alpha=0.6
)

plt.title('Is there a relationship between the total amount paid and the tip amount?')
plt.xlabel('Total Amount')
plt.ylabel('Tip Amount')

cbar = plt.colorbar(orientation='horizontal')
cbar.set_label('Passengers')

plt.show()

5. Creating Bar Charts

5.1 When to Use a Bar Chart

Bar charts are ideal for:

• Comparing categories and ranking them easily
• Analyzing rankings (historical highs and lows at a glance)
• Showing part-to-whole relationships (stacked bar charts)
• Identifying trends and outliers by category

5.2 Demo: Creating Bar Charts

import matplotlib.pyplot as plt
import pandas as pd

full_dataset = pd.concat([pd.read_csv(f'Yellow Taxi Trips - Week {i}.csv') for i in range(1,6)])

trips = full_dataset.groupby(by='Payment_Type', as_index=False)['ID'].count()
trips.rename(columns={'ID': 'Trips'}, inplace=True)

# Vertical bar chart
plt.bar(trips['Payment_Type'], trips['Trips'])
plt.title('Yellow Taxi Trips by Payment Method')
plt.xlabel('Payment Method')
plt.ylabel('Trips')
plt.show()

# Horizontal bar chart
plt.barh(trips['Payment_Type'], trips['Trips'])
plt.show()

Stacked bar chart

weekly_trips = full_dataset.pivot_table(
    values='ID',
    index='Week_No',
    columns='Payment_Type',
    aggfunc='count'
).reset_index()

plt.bar(weekly_trips['Week_No'], weekly_trips['Cash'], label='Cash')
plt.bar(
    weekly_trips['Week_No'],
    weekly_trips['Credit card'],
    bottom=weekly_trips['Cash'],
    label='Credit card'
)
plt.legend()
plt.show()

5.3 Demo: Customizing Bar Charts

colors = ['#00A3FF', '#1B1834']

barchart = plt.bar(
    trips['Payment_Type'],
    trips['Trips'],
    width=0.6,
    color=colors,
    alpha=0.8,
    edgecolor='black',
    linewidth=2,
    linestyle='-'
)

plt.title('Yellow Taxi Trips by Payment Method')
plt.xlabel('Payment Method')
plt.ylabel('Trips')

plt.bar_label(barchart, labels=trips['Trips'])
plt.show()

5.4 Demo: Exporting Figures

# Export to PNG (high resolution)
plt.savefig(
    'barchart.png',
    dpi=300,            # dots per inch (300 = print quality)
    bbox_inches='tight' # avoid clipping
)

# Export to PDF
plt.savefig('barchart.pdf', bbox_inches='tight')

6. Reference Tables

Chart Types

Common Parameters

Line Styles and Markers

Color Specification

7. Reference Diagrams

Visualization Workflow

flowchart LR
    A[Import\ndata] --> B[Prepare\nPandas DataFrame]
    B --> C[Choose\nchart type]
    C --> D[Create\nFigure and Axes]
    D --> E[Plot\ndata]
    E --> F[Customize\ntext, colors, styles]
    F --> G[Display\nplt.show]
    G --> H[Export\nsavefig]

Chart Type Decision

flowchart TD
    Q1{What is\nthe goal?}
    
    Q1 -->|Trend over time| LC["Line Chart\nplt.plot()"]
    Q1 -->|Relationship between 2 variables| SC["Scatter Plot\nplt.scatter()"]
    Q1 -->|Compare categories| Q2{Orientation?}
    Q1 -->|Distribution| H["Histogram\nplt.hist()"]
    Q1 -->|Proportions| P["Pie Chart\nplt.pie()"]
    
    Q2 -->|Vertical| BC["Bar Chart\nplt.bar()"]
    Q2 -->|Horizontal| BH["Horizontal Bar\nplt.barh()"]
    Q2 -->|Composition| SB["Stacked Bar\nplt.bar + bottom="]

Search Terms

matplotlib · python · foundations · data · analysis · engineering · analytics · chart · charts · line · bar · scatter · customizing · parameters · plots · text · api · architecture · color · figures · layer · plot · reference · specification